Cookest
AI & Automation

Agentic Skills

Complete deck of AI agent skills, capabilities, and patterns for the Cookest ecosystem

Agentic Skills

This page defines the complete set of skills an AI agent should have when working on Cookest. Each skill includes its purpose, when to apply it, and how to execute it correctly.

Skill Categories


Code Skills

SK-01: Rust API Development

When: Creating or modifying API endpoints, services, or data models.

Context required:

  • Read api/README.md for endpoint overview
  • Read api/docs/database/SCHEMA.md for database schema
  • Read existing handlers/services in the same domain

Patterns:

Handler (thin) → Service (logic) → Entity (data)

Checklist:

  • Handler validates input using validation/ rules
  • Service contains all business logic
  • Entity uses SeaORM with #[derive(Clone, Debug, DeriveEntityModel)]
  • Errors use AppError from errors.rs
  • Auth middleware applied if endpoint needs authentication
  • Rate limiting configured via governor
  • Subscription tier checked if feature is gated
  • Tests cover happy path + error cases

SK-02: Flutter UI Development

When: Creating or modifying screens, widgets, or state management.

Context required:

  • Read UI/lib/src/shared/theme/ for color/typography constants
  • Read existing feature modules for patterns
  • Check Riverpod providers in the same feature

Patterns:

Screen → Provider (Riverpod) → Repository → API Client (Dio)

Checklist:

  • Use const constructors wherever possible
  • Follow Material 3 theming with sage green (#7A9A65)
  • Riverpod provider per data concern
  • GoRouter for navigation (no manual Navigator.push)
  • Responsive layout considerations
  • Dark mode support via theme tokens
  • Error states handled in UI

SK-03: TypeScript Web/Docs Development

When: Modifying the landing page or documentation site.

Context required:

  • Web: Read web/app/components/ for existing patterns
  • Docs: Read docs/lib/i18n.ts for locale system
  • Read messages/*.json for translation keys

Checklist:

  • Components use TypeScript strict mode
  • Translations use t(locale, key) — no hardcoded strings
  • TailwindCSS for styling
  • Responsive design (mobile-first)
  • Dark mode via theme provider
  • Framer Motion for animations (web only)

SK-04: Python ETL Development

When: Modifying data ingestion pipelines.

Context required:

  • Read etl/README.md for pipeline overview
  • Check api/docs/database/SCHEMA.md for target schema
  • Understand rate limiting configuration

Checklist:

  • Upsert logic (ON CONFLICT) for idempotency
  • Rate limiting respected for external APIs
  • Progress bars via tqdm
  • Error handling with logging (not print)
  • Type hints on all functions

Documentation Skills

SK-05: Content Writing

When: Writing new documentation pages or updating existing ones.

Context required:

  • Read docs/content/docs/contributing/guide.mdx for standards
  • Read existing pages in the same section for voice consistency
  • Check source code for accuracy

Checklist:

  • Frontmatter with title and description
  • H1 matches frontmatter title
  • Max H3 depth (never H4+)
  • Code blocks have language tags
  • Tables for structured data
  • Mermaid for diagrams
  • Callouts for warnings/notes
  • Added to section's meta.json

SK-06: Translation

When: Translating documentation to a new language or updating existing translations.

Context required:

  • Read docs/content/docs/contributing/translations.mdx
  • Read messages/en.json for UI string structure
  • Check language-specific style guide

Checklist:

  • All prose translated, code blocks kept in English
  • Frontmatter translated
  • meta.json updated for the locale
  • JSON message file updated if UI strings changed
  • Community translation banner added (Tier 2 languages)
  • Technical terms kept in English (JWT, API, webhook)

SK-07: Architecture Diagrams

When: Documenting system design, data flows, or component relationships.

Tool: Mermaid (rendered natively by Fumadocs).

Diagram types:

TypeUse Case
graph TDComponent architecture, data flow
sequenceDiagramAuth flows, request/response
erDiagramDatabase schema
flowchartDecision trees, pipelines

SK-08: API Reference Documentation

When: Documenting new or changed API endpoints.

Structure per endpoint:

## Endpoint Name

| Method | Path | Auth | Tier |
|--------|------|------|------|
| POST | `/api/v1/resource` | Bearer | Pro |

### Request Body
{code block with JSON example}

### Response
{code block with JSON example}

### Error Codes
{table of possible errors}

Operations Skills

SK-09: Code Review

When: Reviewing PRs from human or AI contributors.

Focus areas:

  1. Correctness — Does the code do what it claims?
  2. Security — No secrets, proper auth, input validation
  3. Performance — No N+1 queries, appropriate indexing
  4. Consistency — Follows project conventions
  5. Documentation — Changes reflected in docs

SK-10: PR Management

When: Creating, updating, or merging pull requests.

Commit conventions:

<type>(<scope>): <description>

PR title: Same format as commit messages.

PR body: Must include What, Why, How, Testing sections.

SK-11: CI/CD Pipeline

When: Adding or modifying automated workflows.

Available pipelines:

  • auto-docs.yml — Detects code changes, suggests doc updates
  • translation-check.yml — Reports translation coverage

SK-12: Dependency Audit

When: Reviewing or updating project dependencies.

Per-language tools:

LanguageAudit Command
Rustcargo audit
Dartflutter pub outdated
Node.jsnpm audit
Pythonpip-audit

Skill Application Matrix

SkillapiUIwebdocsetl
SK-01 Rust API
SK-02 Flutter UI
SK-03 TypeScript
SK-04 Python ETL
SK-05 Content Writing
SK-06 Translation
SK-07 Diagrams
SK-08 API Reference
SK-09 Code Review
SK-10 PR Management
SK-11 CI/CD
SK-12 Dependency Audit

On this page